Chris Lehane, OpenAI’s global policy chief, confirmed Tuesday that the company has been collaborating with rivals Anthropic and Google DeepMind on AI safety protocols for several weeks. This admission validates earlier reports from Bloomberg and The Information suggesting a covert alignment among the industry’s biggest players. Lehane made the statement while in Washington, engaging with U.S. lawmakers to address catastrophic risks associated with frontier models. The timing is no accident. It follows Anthropic CEO Dario Amodei’s recent essay urging the sector to decelerate development to prevent existential threats. Major figures like Sam Altman, Demis Hassabis, and Elon Musk have publicly backed Amodei’s call, signaling a rare moment of unity in a typically cutthroat market.
Antitrust concerns loom large over these private discussions. Coordination between competitors can easily be interpreted as price-fixing or market suppression, legal pitfalls that usually keep corporate lawyers awake at night. Amodei proposed a narrow government waiver to shield these safety collaborations from legal scrutiny, but Lehane dismissed the need for such protection. He argued the firms are operating within existing legal frameworks. This stance contradicts reports that Altman told staff a new industry standards body would need to form without direct U.S. government support. The disconnect suggests internal uncertainty about how far these alliances can stretch before regulators intervene. Google’s Hassabis has previously advocated for a powerful AI watchdog with authority to screen advanced models, a move that would effectively institutionalize these informal talks.
Political headwinds remain strong. Former President Trump has labeled safety concerns a hoax, arguing that regulatory slowdowns only benefit China. His advisor David Sacks, an investor with deep ties to the tech sector, echoed this sentiment by downplaying existential risks. Despite this opposition, OpenAI is pushing for specific legislative measures. Lehane expressed support for a provision in the FRONTIER Act that mandates independent verification organizations access to top labs. This move toward third-party auditing represents a tangible shift from voluntary guidelines to enforced transparency. Whether Washington will accept this industry-led oversight model remains unclear, especially given the current political climate’s skepticism toward tech regulation. The industry is betting that self-policing looks better than federal intervention, but that bet carries significant legal and reputational risk.
